Filter Forge - 3.006 Issues

Sphinx.

Filter Optimizer

Posts: 1750
Filters: 39

Strange things happen in 3.006 - I'm sorry I can't be more specific - I simply can't find any pattern to it. It is related to bad floating point numbers though..

I did send you a mail about this already - no answers..

Take a look at the attached filter - open the editor, find the Extract HSB that goes to a Tone Curve ("TC") - now try connecting the TC to the Blend that the HSB goes to.. I dropped a checkbox nearby with the text ("Connect TC to Blend!").

This causes strange things to happen to the result..

Fast Collage.ffxml

Njyldgarkn init/prepare process x no. of render blocks x no. of passes

Posted: November 28, 2011 5:06 am

Details E-Mail

Sphinx.

Filter Optimizer

Posts: 1750
Filters: 39

Before problem shows:

Njyldgarkn init/prepare process x no. of render blocks x no. of passes

Posted: November 28, 2011 5:10 am

Details E-Mail

Sphinx.

Filter Optimizer

Posts: 1750
Filters: 39

After problem shows:

Njyldgarkn init/prepare process x no. of render blocks x no. of passes

Posted: November 28, 2011 5:11 am

Details E-Mail

Sphinx.

Filter Optimizer

Posts: 1750
Filters: 39

And some strange FP values..

Njyldgarkn init/prepare process x no. of render blocks x no. of passes

Posted: November 28, 2011 5:12 am

Details E-Mail

Sphinx.

Filter Optimizer

Posts: 1750
Filters: 39

The bomber might be the culprit here, because it has been involved in all encounters of this strange problem; the error just never shows immediately after the bomber..

Njyldgarkn init/prepare process x no. of render blocks x no. of passes

Posted: November 28, 2011 5:14 am

Details E-Mail

Sphinx.

Filter Optimizer

Posts: 1750
Filters: 39

Another buggy screenshot, notice the blend settings :-S While very annoying, this one does create a rather interesting output smile:-)

Njyldgarkn init/prepare process x no. of render blocks x no. of passes

Posted: November 28, 2011 8:55 am

Details E-Mail

saurabhgayali

Saurabh Gayali

Posts: 183
Filters: 84

one problem or suggestion in filter.
though we increase repeat the image scale almost remains constant

saurabh.gayali@gmail.com

Posted: December 7, 2011 11:50 am

Details E-Mail

BenBeckwith

Posts: 136
Filters: 8

Man I've been running into these issues all week and just came by to post about it, so I'm glad to see this.

Here's a simple test case I found that shows weird results.

Bomber Blend Bug.ffxml

Posted: December 15, 2011 10:11 pm

Details E-Mail

BenBeckwith

Posts: 136
Filters: 8

And here's an image. The only part that isn't default is the last blend biased 100% to the high pass input in order to show the undesired effect. If you bypass the first blend then all works well. If you make a copy of the bomber and make it so one goes into the blend and the other into the high pass then it will also work fine. So it's just very weird, but hopefully that helps narrow it.

Posted: December 15, 2011 10:20 pm

Details E-Mail

Sphinx.

Filter Optimizer

Posts: 1750
Filters: 39

This is very annoying - but even more so is that FF has not responded at all - I contacted them by e-mail also. smile:|

Njyldgarkn init/prepare process x no. of render blocks x no. of passes

Posted: December 16, 2011 10:56 am

Details E-Mail

GMM
Moderator

Filter Forge, Inc

Posts: 3491

This looks like another manifestation of the Subtle Bomber Bug (case 1, case 2, case 3). Dmitry aka Egret claims he has fixed the bug â€“ please wait for the next program update smile:)

Posted: January 30, 2012 8:12 am

Details E-Mail

Mardar

Graphics Junkie

Posts: 688
Filters: 61

Quote
Dmitry aka Egret claims he has fixed the bug â€“ please wait for the next program update

This is great news. I was working on another filter last night and again noticed weird markings in the bomber preview but they didn't show up in the final result. So I was hoping this could be resolved soon. smile;)

Posted: January 30, 2012 8:58 am

Details E-Mail

SpaceRay

SpaceRay

Posts: 12298
Filters: 35

Quote
GMM wrote: This looks like another manifestation of the Subtle Bomber Bug (case 1, case 2, case 3). Dmitry aka Egret claims he has fixed the bug â€“ please wait for the next program update

This was in 3.006 and now we are in 3.008, and I do not know if this has been fixed yet or not, anyone knows?

Thanks

Posted: June 4, 2012 2:24 pm

Details E-Mail

Sphinx.

Filter Optimizer

Posts: 1750
Filters: 39

It has been fixed (tested in 3.008, probably fixed in 3.007)

Njyldgarkn init/prepare process x no. of render blocks x no. of passes

Posted: June 7, 2012 10:48 am

Details E-Mail